acappiello/msqc

Quantum chemistry

msqc's Issues

Consolidated user-specific options

It shouldn't be necessary for each user to change things like dataset locations or scratch directories repeatedly.

Idea: global config struct/class.

Add documentation

Documentation is extremely inconsistent throughout the project. I propose we use README files for big picture, high level documentation and the matlab files for more detailed documentation.

For matlab functions, I propose a format like:

function bar = fun(foo)
% Description:
%   A brief overview of what the function does.
%
% Input:
%   foo:    Some data.
%
% Output:
%   bar:   Some result.
%
% Example (optional):
%   x = blah;
%   var = fun(x);

Overly restrictive Environment.compare

This code will not detect identical permutations.

For example:

%% Bug demonstration.
x = Environment;
x.ncharge = 2;
x.rho = [10 15];
x.r = [1 2; 3 4; 5 6];

y = Environment;
y.ncharge = 2;
y.rho = [15 10];
y.r = [2 1; 4 3; 6 5];

disp(x.compare(y));

However, I can't imagine that this is a common case.

Regression testing

There is currently no straightforward way to test changes in the code for bugs.

Dipole Data

Data for Fragment.dipole and Fragment.dipoleEnv is missing from generated datasets.

The data does not appear to exist in the checkpoint file.

Scripts are clutting the root directory

These should be organized and moved into folders.

Legacy code

Decide whether to drop support (a.k.a. remove) code for we very likely won't revisit, such as interpolation.

Dependent Mixers

There are no checks in place to ensure that multiple mixers aren't redundant, causing them to push and pull without any effect.